A Peptide Amphiphile Organogelator of Polar Organic Solvents

A peptide amphiphile is reported, that gelates a range of polar organic solvents including acetonitrile/water, N,N-dimethylformamide and acetone, in a process dictated by β-sheet interactions and facilitated by the presence of an alkyl chain. Similarities with previously reported peptide amphiphile hydrogelators indicate analogous underlying mechanisms of gelation and structure-property relationships, suggesting that peptide amphiphile organogel design may be predictably based on hydrogel precedents.

with sonication it dissolves in mixtures of acetonitrile and water, which were therefore used to study gelation. The minimum gel concentration (MGC) in a mixture of 50% acetonitrile/water (v/v) was determined to be 0.06 wt.% (0.64 mM). Thereafter, 0.20 wt.% (2.14 mM) solutions of VVVGHVVV-C 8 in 10, 20, 30, 40 and 50% acetonitrile/ water were used to study gelation, at room temperature. By the inversion test, gelation occurred after 10, 8, 8 and 5 days, and in less than 1 day, for these solutions respectively, showing that more acetonitrile resulted in faster gelation. Increasing the concentration of VVVGHVVV-C 8 also sped up gelation with 0.8 wt.% (8.56 mM) solutions in 50:50 acetonitrile/water forming gels almost immediately after initial sonication. Since the 0.20 wt.% 50% solution formed most readily, it was used as the focus for further studies. The unalkylated VVVGHVVV was also soluble in acetonitrile/water mixtures, however these solutions did not gelate.
AFM of the gel of VVVGHVVV-C 8 (Fig. 2a) shows short fibrils ranging in height from 8-12 nm. These are characterized by a valley along the length of each fibril (Fig. 2b), suggesting collapse of the core probably as a consequence of solvent loss during sample preparation. Strain sweep rheological experiments performed on the gel (Fig. S8) formed from a 0.20 wt.% solution at a constant frequency of 1 Hz, showed that the mechanical properties deviated from the linear viscoelastic region at less than 1% strain, and the cross-over point from gel to liquid (where the storage modulus G′ = the loss modulus G″ ) occurred at 3% strain. An oscillatory frequency sweep performed (Fig. S9) at a constant 0.2% strain demonstrated that the gel is frequency independent over the range studied, with a stiffness (approximated by G′) of 1.5 × 10 5 Pa. VVVGHVVV-C 8 therefore forms a very stiff but brittle gel. A thixotropic test, with application of a low strain (0.2%, 200 seconds) followed by a large strain (100%, 30 seconds), acts to completely liquefy the gel network, and upon repetition of this strain sequence, shows that the recovery of the mechanical properties of the gel is fast but incomplete (Fig. 3).
The CD spectrum of the gel (Fig. 4) shows a minimum at 218.3 nm, characteristic of β -sheet secondary structure 31 . The spectrum of a corresponding solution of VVVGHVVV-C 8 before gelation occurs shows a minimum at 200.9 nm with a shoulder at ∼ 215 nm, indicating a mixture of random coil/disordered and β -sheet structures, respectively. The differences between these spectra show that gelation involves rearrangement of these disordered structures into a β -sheet assembly. Even though the unalkylated VVVGHVVV does not form a gel, its CD spectrum also shows an intense minimum at 214.1 nm characteristic of β -sheet formation, demonstrating that gelation is not solely determined by interactions between peptide chains and the alkyl moiety of VVVGHVVV-C 8 is key to gel formation.
The structural information provided by CD spectroscopy is mirrored in results obtained using FT-IR spectroscopy. The gel of VVVGHVVV-C 8 , prepared in 50% acetonitrile/D 2 O instead of acetonitrile/water so as to  eliminate interference from H 2 O 32 , showed an amide I carbonyl signal at 1624 cm −1 corresponding to a β -sheet structure 33 . The signal at 1626 cm −1 in the corresponding spectrum of the unalkylated VVVGHVVV shows a similar structure, again demonstrating that gelation is not only dependent on peptide secondary structure, and that the alkyl moiety of VVVGHVVV-C 8 is required. The carbonyl signals in the spectra of both the gel of VVVGHVVV-C 8 and the solution of VVVGHVVV are each red-shifted by approximately 5 cm −1 , relative to those of the corresponding powders, which appear at 1630 and 1631 cm −1 , respectively. This provides tentative evidence that the interactions between the peptide chains in the gel of VVVGHVVV-C 8 are similar to those in the solution of VVVGHVVV.
Considering solvents other than acetonitrile/water, VVVGHVVV-C 8 did not dissolve in ethyl acetate, diethyl ether or hexane despite heating to reflux temperatures and sonicating for up to one hour. VVVGHVVV-C 8 dissolved in dimethyl sulfoxide, methanol, ethanol, isopropanol, chloroform, dichloromethane and toluene but did not form gels from those solutions, and dissolved in DMF and acetone to give solutions that form gels with MGCs of 0.06 wt.% in each case. No gelation of VVVGHVVV-C 8 was observed in methanol even though the CD spectrum shows that strong β -sheets are formed (Fig. 4). AFM of the DMF gel ( Fig. 5a) shows fibers that are much longer (several microns) and more bundled (with a larger diameter of 10-20 nm) than the fibrils of the acetonitrile/ water gel. In the case of acetone (Fig. 5b), strong fiber bundling is observed, precluding measurements of the fiber diameter.
Circular dichroism measurements of the acetone and DMF organogels were not possible, because of strong solvent absorption obscuring any secondary structure information. FT-IR spectroscopy of the DMF gel was not possible as the carbonyl signal is masked, but FT-IR of the acetone organogel shows a carbonyl stretch at ∼1629 cm −1 (Fig. S10 and S11), representative of a β -sheet structure. This suggests that the secondary structure of VVVGHVVV-C 8 is preserved when the gelating solvent is changed. Rheology experiments (Figs S14 and S15) on the VVVGHVVV-C 8 DMF organogel show that it is two orders of magnitude weaker than the corresponding water/acetonitrile gel, but more amenable towards oscillatory strain, with no cross-over point observed. No previously reported DMF organogels [34][35][36] approach the gel stiffness (taking into account relative wt.%) or minimum gel concentrations observed in this study, making it unique.  Basis of Gel Formation. As previously discussed, VVVGHVVV-C 8 is structurally similar to PAs that form hydrogels, in that it comprises an alkyl chain, a peptide region and a charged head group. Even so, it is unsuitable for hydrogel formation because it does not dissolve in water, presumably because its head group is less polar than, for example, that of EEEAAAVVV-C 16 . It is however, soluble in organic solvents and that provided the opportunity to explore its organogel behavior and observe acetone, DMF and acetonitrile/water gels. Organogelators that comprise a peptide component are already known but they do not have the other components of PAs [37][38][39] . Some do form β -sheets 35,40,41 , but gelation generally depends on π -π -interactions 36,42,43 .
CD and FT-IR spectra show that gelation of VVVGHVVV-C 8 involves β -sheet formation. Since VVVGHVVV also forms β -sheets without forming gels however, it is clear that these interactions are not the sole basis for extended fiber formation, and both the alkyl chain and peptide component are required. With hydrogelating PAs it has also been observed that an alkyl chain of more than six carbons is necessary for gelation to occur 44 . Although the details of the gelation of VVVGHVVV-C 8 are not delineated by our studies, the analogy with PAs such as EEEAAAVVV-C 16 suggests a similar gelation mechanism despite the difference in solvent; that is micelle-type hydrophobic aggregation of alkyl chains which propagates down the fiber axis, then reinforced by peptide β -sheet interactions extending along the length of the fiber.
The stiffness of the acetonitrile/water gel of VVVGHVVV-C 8 (1.5 × 10 5 Pa) is higher than that of similar PAs such as EEEAAAVVV-C 16 (< 10 4 Pa). This is consistent with previous findings that increasing the proportion of valines in the β -sheet forming region of a PA results in an increase in gel stiffness 26 . With PAs such as EEEAAAVVV-C 16 , the β -sheet signal in the CD spectrum is red-shifted by a magnitude that inversely correlates with gel stiffness, and this has been attributed to the extent of twisting of the sheets, i.e., stiffer gels involve stronger peptide interactions with the least twisting 26 . Again, the gel of VVVGHVVV-C 8 fits this trend. The corresponding signal is red-shifted by only 2.3 nm, consistent with a very stiff gel.

Conclusion
In summary, VVVGHVVV-C 8 is a peptide amphiphile that gelates non-aqueous solvents. Assembly occurs in acetonitrile/water, DMF and acetone via β -sheet interactions, facilitated by the presence of the alkyl chain. Analogies with previously reported PA hydrogelators suggest that the underlying mechanism of gelation is the same despite the difference in solvent, and that structure-property relationships identified with these PAs also hold true. This study therefore indicates that the existing extensive literature that reports on PA hydrogels may be drawn upon to usefully inform PA organogel retro-design and predictability, for assembly in a range of solvents.

Rheological Measurements.
Peptide solutions (0.20 wt.%, 2.14 mM) were made by dissolving in 50% acetonitrile/water with sonication. An aliquot (550 μ L) of this solution was then cast onto an Anton Paar MCR 302 rheometer (parallel plate geometry, 25 mm diameter, 1 mm gap) and left to gel over the course of 4 h. A solvent trap and Peltier hood was used to prevent any evaporation during gelation. Frequency sweeps were performed from 0.1-10 Hz, at a constant strain of 0.2% and three repeats performed on separate samples. Strain sweeps were performed from 0.1-100% at a constant frequency of 1 Hz. Circular Dichroism Spectroscopy. Solutions of VVVGHVVV-C 8 (0.20 wt.%, 2.14 mM) were made by dissolving in 50% acetonitrile/water with sonication. These were then diluted by serial dilution with 50% acetonitrile/water until HT < 1,000 across the sample spectra range of 190-260 nm. Spectrometry was performed using a Chirascan CD spectrometer and analyzed using Pro-data software. Samples were measured in a 10 mm Hellma QS cuvette using a 0.3 nm step and 1.5 s per point at 20 °C. The baseline in each case was a blank solution of 50% acetonitrile/water. The average spectra of 5x scans were taken at 27.5 μ M, 20.6 μ M, 15.5 μ M and 11.6 μ M and their molar ellipticities then averaged to give the CD spectrum of VVVGHVVV-C 8 before gelation.
The 2.14 mM solution of VVVGHVVV-C 8 was left to gel overnight then shaken to give a solution that was diluted by serial dilution with 50% acetonitrile/water. The average spectra of 5x scans were taken at 27.5 μ M, 20.6 μ M, 15.5 μ M and 11.6 μ M and their molar ellipticities then averaged to give the CD spectrum of VVVGHVVV-C 8 after gelation.
A concentrated solution of VVVGHVVV was made in 50% acetonitrile/water and diluted by serial dilutions with 50% acetonitrile/water. The average spectra of 5x scans were taken at 29.06 μ M, 21.80 μ M, 16.35 μ M and 12.26 μ M and their molar ellipticities averaged to give the CD spectrum of VVVGHVVV.
A 2.14 mM solution of VVVGHVVV-C 8 in methanol was made, then diluted by serial dilutions. Spectra in this solvent were measured across a range of 195-260 nm because of the high absorbance of methanol at lower wavelengths. The average spectra of 5x scans were taken at 31.25 μ M, 15.6 μ M and 7.8 μ M and their molar ellipticities then averaged to the CD spectrum of VVVGHVVV-C 8 in methanol.